+- The Voynich Ninja (https://www.voynich.ninja)
+-- Forum: Voynich Research (https://www.voynich.ninja/forum-27.html)
+--- Forum: Analysis of the text (https://www.voynich.ninja/forum-41.html)
+--- Thread: Deconstructing the VMS : A positionless ledger (part 1) (/thread-6040.html)
RE: Deconstructing the VMS : A positionless ledger (part 1) - Dunsel - 25-08-2026
(25-08-2026, 10:10 AM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.Apologies for misunderstanding the nature of your ledger.
But if the ledger has cycles, it accepts an infinite set of word types, no? So you have a separate limit on the length of the words? Or you generate text without spaces, and insert spaces by a separate method?
Does your ledger accept words with two gallows, or with an EVA [dlrs] between two benches {ch,sh}?
No apologies needed.
The ledger itself doesn't have cycles and it cannot 'accept' an infinite set of word types but it could generate an infinitely long string. The ledger is not itself a generator so it doesn't even recognize spaces. The scribe is the generator. The scribe has to decide when a word is long enough and add a suffix.
The ledger provides the combinatorial rules needed to build words. It tells the scribe that if they have the letter A, here are the legal midfixes and suffixes that can follow that letter. It would tell the scribe that they could create the word daiin but they could not create dagrgy.
The ledger itself is nothing more than a rule book for creating Voynich words. It doesn't tell you how to create the line or the page.
Quote:Ok. But I don't think there was ever a question of whether the hypothetical five scribes used different vocabularies. The claim was that there were only two "languages", and Herbal was divided among them.
It would make more sense to investigate vocabulary sharing among the sections, treating Herbal-A and Herbal-B as two sections. It is expected that there will be differences in vocabulary due to different subject matters.
Actually, I think I can demonstrate that there were not 5 or even 2 different vocabularies. And I will debate whether there are two Currier languages. In my next post, I'll try to demonstrate that all 5 scribes and both languages are nothing more than scribal preferences using the same system.
Quote:It accepted 89.3% of [scribes 2-5] tokens and 77.6% of their types.
Quote:Those numbers seem rather low. But low numbers are expected because of the small size of scribe-1 sample and the variety of topics.
Yea, that number is actually a bit misleading. Scribe 2 used all of the ledger that scribe 1 had available and added 3 new midfix transitions: C→S, R→CH, and S→E.
The other difference between the two scribes is preferences. If you remember my work on the bigram <ed> and how scribe 2 used it explosively, that's the big visible difference Currier identified. Scribe 1 already had <ed> available to use but choose to not use it often. Scribe 2 moved that transition from rare to extremely common.
So scribe 2 did not invent a new grammar. It reused Scribe 1’s ledger, promoted some rare options into normal use, added a few new transitions, and changed which constructions were favored.
And... here's some more foreshadowing. Scribe 1 may not have been the first scribe to start writing the Voynich, an assumption most all of us have. When I look at how the ledger changed and expanded for each scribe, the tentative scribal chronological order appears to be 5 → 4 → 3 → 1 → 2. That being said, scribes 4 and 5 have relatively small samples so that's VERY tentative.
RE: Deconstructing the VMS : A positionless ledger (part 1) - oshfdk - 25-08-2026
(25-08-2026, 12:55 PM)Dunsel Wrote: You are not allowed to view links. Register or Login to view.The ledger itself doesn't have cycles and it cannot 'accept' an infinite set of word types but it could generate an infinitely long string.
I'm completely lost. How is it possible to generate an infinitely long string without cycles?
Also, could you clarify if various claims and numbers in this thread are verified by yourself or just produced by an LLM? I mean, when you write "as a positionless conditional-ledger it scores 479,627 Nbits with gallows retained and 416,094 with gallows stripped", is this something you personally verified or just something produced by an LLM generated piece of code that you cannot analyze?
RE: Deconstructing the VMS : A positionless ledger (part 1) - nablator - 25-08-2026
(25-08-2026, 01:36 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.I'm completely lost. How is it possible to generate an infinitely long string without cycles?
By not repeating the same sequence ad infinitum. No problem.
RE: Deconstructing the VMS : A positionless ledger (part 1) - oshfdk - 25-08-2026
(25-08-2026, 01:51 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.By not repeating the same sequence ad infinitum. No problem.
Ok, I think I understand. I guess if irrational numbers can pull this off, why not. I'm then confused about the meaning of 7.82 billion possible word forms. If infinite strings are allowed, there should be infinitely many forms, since an infinite string can be terminated at any moment. Maybe it's possible word forms up to a certain length?
RE: Deconstructing the VMS : A positionless ledger (part 1) - nablator - 25-08-2026
(25-08-2026, 02:08 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.Maybe it's possible word forms up to a certain length?
I don't know.
My preference would be to simplify the ledger enough to prevent infinite strings, by removing the rarest transitions. For example d after n does occur but a more parsimonious explanation is that a space is missing. These compound words would fail the check but I see that as an improvement, because spaces cannot be trusted (especially in transliterations) and as long as the subdivision does not create any new hapax: not sure if that's possible. Anyway, the "no hapax" rule is probably not really needed if the rarest transitions are eliminated, it will take care of possible scribal errors and misreadings.
RE: Deconstructing the VMS : A positionless ledger (part 1) - Dunsel - 25-08-2026
(25-08-2026, 01:36 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.I'm completely lost. How is it possible to generate an infinitely long string without cycles?
Also, could you clarify if various claims and numbers in this thread are verified by yourself or just produced by an LLM? I mean, when you write "as a positionless conditional-ledger it scores 479,627 Nbits with gallows retained and 416,094 with gallows stripped", is this something you personally verified or just something produced by an LLM generated piece of code that you cannot analyze?
Again, read how the ledger works. If you choose, it can keep right on creating bigrams until YOU decide to end the word with a suffix. It is NOT A GENERATOR!
And that particular piece of code, no I didn't go over it. Everything else I check. I gave you the code. Do with it as you wish. Or, if you'd just rather I say, I dunno... then so be it. I dunno. And honestly, it's the wrong comparison for what I'm presenting anyway.
RE: Deconstructing the VMS : A positionless ledger (part 1) - Dunsel - 25-08-2026
(25-08-2026, 02:08 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.(25-08-2026, 01:51 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.By not repeating the same sequence ad infinitum. No problem.
Ok, I think I understand. I guess if irrational numbers can pull this off, why not. I'm then confused about the meaning of 7.82 billion possible word forms. If infinite strings are allowed, there should be infinitely many forms, since an infinite string can be terminated at any moment. Maybe it's possible word forms up to a certain length?
That was based on the dictionary it was forcing it into in order to try to give you a comparison.
RE: Deconstructing the VMS : A positionless ledger (part 1) - Dunsel - 25-08-2026
(25-08-2026, 02:29 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.(25-08-2026, 02:08 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.Maybe it's possible word forms up to a certain length?
I don't know.
My preference would be to simplify the ledger enough to prevent infinite strings, by removing the rarest transitions. For example d after n does occur but a more parsimonious explanation is that a space is missing. These compound words would fail the check but I see that as an improvement, because spaces cannot be trusted (especially in transliterations) and as long as the subdivision does not create any new hapax: not sure if that's possible. Anyway, the "no hapax" rule is probably not really needed if the rarest transitions are eliminated, it will take care of possible scribal errors and misreadings.
Words formed until the scribe wanted them to stop. If the Voynich is copy and mutate the ledger wasn't used to create words. It was used to keep the scribes on track with valid transitions when they did modify a word. If you copy a word, you don't need a ledger, just copy it.
And yes, a missing space is entirely possible and, very likely probable. However, those would be subject to interpretation. The ledger can accommodate ANY language and ANY transcription. If you think one transcription has bad spaces, try another. But those spaces are subjective. If I did not include d after n then someone would complain that the ledger is incomplete because I didn't include them. That's why I've split them into common and rare.
But you're free to use this information however you wish. Please, experiment. Find new things.
RE: Deconstructing the VMS : A positionless ledger (part 1) - Dunsel - 25-08-2026
(25-08-2026, 01:51 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.(25-08-2026, 01:36 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.I'm completely lost. How is it possible to generate an infinitely long string without cycles?
By not repeating the same sequence ad infinitum. No problem.
Exactly. If you keep doing prefix->midfix->prefix->midfix it will let you keep adding letters to the word until you select prefix->suffix
RE: Deconstructing the VMS : A positionless ledger (part 1) - Jorge_Stolfi - 26-08-2026
(25-08-2026, 01:51 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.You two seem to be talking about different things.(25-08-2026, 01:36 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.I'm completely lost. How is it possible to generate an infinitely long string without cycles?By not repeating the same sequence ad infinitum. No problem.
If the ledger can generate infinitely many strings, it must have a cycle. That's because a set of infinitely many words must have arbitrarily long words. If a finite state automaton with N states generates a word with N or more letters, it must go twice or more through some state. That means the automaton has a cycle. Then traversing that cycle M times generates a word that contains a substring that is repeated M times, for any M.
On the other hand, if the alphabet has N letters, it is possible to create a word with ~N^K letters that does not repeat any substring of K letters. These words are called "De Bruijn sequences". For example, with 10 letters one can create a text of a million letters that does not repeat any substring of 6 or more letters.
All the best, --stolfi